Azure Machine Learning service was configured with a Load Balancer & Premium SSD Managed Disks (Cost Issue)

Question

I create an Azure Machine Learning service and it was configured with a Load Balancer & Premium SSD Managed Disks. Is this configuration correct?

I chose a Public Endpoint for the ML service. Is this the reason for the Load Balancer?

Is there a way for me to change the "Premium SSD Managed Disks" to a lower-cost storage?

Answer 1

@Terrence Rideau The load balancer is a sub resource that is created when you create the compute instance or cluster. These resources are charged unless the compute instance or cluster is completely deleted. If the compute is in stopped state the resources might still accrue costs. Please check this section of the documentation for details. The endpoints should not be the reason for billing under load balancer. If you have any questions regarding billing you can always raise a azure support case from azure portal to seek clarification.

The storage that is created with the compute instances chosen from Azure ML portal ml.azure.com cannot be changed. If you would like to use a custom compute type with lower cost storage then you need to use attached compute which is actually created from Azure portal and then added to Azure ML compute(attached) to be used for your experiments. I hope this helps!!

If an answer is helpful, please click on or upvote which might help other community members reading this thread.